Space vector modulation

Under the name space vector modulation (english space vector modulation ( SVM, SVPWM ), or English space vector control ) is understood in the power electronics, a method for control of rotating electrical machines based on the pulse width modulation.

Based on this type of modulation, it is possible, a multi-phase alternating current system replicate electronically, as needed for the operation of three-phase machines. The space vector representation satisfy two sizes, the angle of the space vector and the amount or real and imaginary parts, to specify the flux density distribution in the machine.

• 2.1 implementation
• 2.2 Application

Operation

Prerequisite

If a three-phase system to be simulated, one requires a half-bridge ( HB1, HB2, HB3 ), each of the three phases. Thus, the output voltage of the phase (U, V, W) can be placed on both the positive and the negative DC potential (UDC ). As the circuit is shown already, only one switch may be closed at each half- bridge, otherwise the intermediate circuit voltage is short-circuited. For the further considerations it is assumed that, for each half bridge in each case a switch is closed. Is thus at any time in each phase of a specific potential before. It follows that each half-bridge can take two states. In the first state of the upper switch (state " 1") and in the second state of the low-side switch ( "0" ) is closed.

Basic voltage space vector

Each half- bridge can take two different switch positions. Since three half-bridges are necessary for a three-phase alternating current system, arise from the fact possible switch positions and thus 8 switching states. This results in each switch position a different constellation voltage between the phases and therefore a different voltage space vector. The two switch positions, where either all three upper or all lower three switches are closed are exceptions These switch positions are short-circuited all three phases. Thus, no voltage can be measured between the phases. These two voltage vectors are referred to as zero voltage space vector. This results in six active and two passive voltage space vector can be represented. In the following table are respectively the concatenated output voltages of the 8 switch positions that can occur mapped.

Modulation

Each voltage space vector generated also in a three-phase machine with a specific orientation of the flux density distribution. To a three-phase machine to be able to commute continuous ( sinusoidal), are inadequate for the 6 active basic voltage space vector, since voltage space vector must be connected to the machine with arbitrary angles and amounts.

To accomplish this, the basic principle of pulse width modulation. If you want such as a voltage space vector (Ua) spend the exactly half the angle of the voltage space vector U1 and U2 has, this can be realized by the voltage space vector U1 is alternately output to the voltage space vector U2. The length of time which must be applied each voltage space vector, depends on the switching frequency of the modulation. The decisive factor for the resulting voltage space vector is only the ratio of the two times. In the given example the two times thus need to be exactly the same length selected to obtain the desired voltage space vector. Due to the low-pass effect of the stator windings is obtained in the machine and thus the desired mean current space vector, the desired orientation of the magnetic flux density.

The control logic must therefore consider the time being, in which the six sectors of the desired voltage space vector and is alternately output the two affected basic voltage space vector. The ratio of the times, which must be applied to each of the two voltage space vector, arises from the relative angle of the desired voltage space vector with respect to the angles of the affected basic voltage space vector.

So far, it has been described as arbitrary voltage space vector can be issued, each with the maximum amount. However, for the commutation of an induction machine, it is imperative the amplitude of the output voltage, ie, can choose arbitrarily the amount of voltage space vector. To realize this, the two zero voltage space vector are needed. If you now wish, for example, the voltage space vector Ub spend, the ratio of the output times of the voltage space vector U1 and U2, as in the previous example, be exactly the same. In order to reduce the magnitude of the resultant voltage space vector now, an additional time is introduced, in which a zero voltage space vector is output. The magnitude of the resultant voltage space vector therefore depends on the ratio of on time of the active voltage space vector, and the ON time of the zero voltage space vector.

For output any voltage space vector so each switching period is divided into three periods. In two of these periods, the two active voltage space vector and output a passive voltage space vector in the third period. The three participating voltage space vector (and thus switch positions ) are thus pulse- width modulated.

Optimization and over-modulation

For outputting any voltage space vector that is a zero voltage space vector is always necessary. Since the machine current in the stator winding of an induction machine is the smoother, the higher the switching frequency is chosen, it provides itself, the time that must be spent of zero voltage space vector to be halved. This ensures that for each switching period, a plurality of switching operations to be carried out and thus the frequency is increased. Thus, it is outputted at the start of each switching period of the first half of the zero-voltage space vector, and then the first active voltage space vector, the second half of the zero- voltage space vector, and finally the second active voltage space vector.

In order to minimize switching losses, the control logic may be configured to output the each of the two lower zero voltage space vector. The lower zero voltage space vector is the one which must be for at the respective switch position at least switch switched.

Thus the output voltage remains linked sinusoidally upon rotation of the voltage space vector, of each of voltage space vector is allowed to move only in the space vector diagram drawn in the circuit. For special applications (short-term higher torque ) will modulate the output voltage. The voltage space vector is no longer moved in this case on a circular path, but - in extreme cases - along the drawn in the space vector diagram of hexagon. It should be noted, however, that the resulting output voltage is no longer sinusoidal, but overlaid with harmonics. This results in an induction machine higher losses.

Practice

Implementation

Space vector modulation is implemented usually with microcontrollers or digital signal processors. The technical implementation can be done in software or in hardware, depending on the type of processor used. Special controllers have been in the hardware a the corresponding switching logic, which assumes the output the required voltage space vector. In this case, the user must simply pass the desired voltage space vector and a register of the hardware portion of the controller providing the correct modulation, in order to obtain the desired voltage space vector.

Application

Common applications for the space vector modulation in frequency converters for induction machines, and partly in power converters for brushless DC motors ( PMSM specifically ), which are very similar to a frequency converter. Specifically, the rotating field machines commutate power converters using the field-oriented control, space vector modulation plays an important role, as even the control is carried out based on the space vector representation.

Another application are special three- phase rectifier such as the Vienna rectifier, where this technique is used for reduction of harmonics.